motion.adaptive-feed

Hello Guys,

by using MUX4, we managed to set the proper value.
It is a pity that we cannot show this simultaneous in a bar on the pyvcp panel.
LinuxCNC complains==>Only one HAL pin available .
Is there a way around?

many thanks upfront.

Thank you andypugh for giving attention to our post.

No error messages whatsoever, MUX4, MUX8 and the bar in PYVCP all working fine.

We are talking Wire EDM.
When eroding the gap voltage between the wire and the workpiece is inverse proportional towards the distance between both.
I mean, when erosion takes place, the nearer the wire gets to the workpiece, the lower the generator voltage becomes.
We detect this voltage in a "gap detector" an put out a 3 bit signal over an IO board towards HAL.
This detection goes through MUX8 and gives an adaptive feed to the wire.
This is what the creator of the motion.adaptive-feed meant in creating this pin.
So far, so good.
Our handicap, at this moment, is that the decision making of the gap detector is that fast, that it is difficult to follow with the eye.
A visualisation bar in PYVCP would help a lot. The equivalent is present in all Wire EDM's like AGIE, Charmilles and GF+.
This bar helps the operator in the decision making to hold or change the generator settings.

In LinuxCNC terms:
by using MUX8 we occupy the hal pin motion.adaptive-feed. Once used, it can't be used once again to visualise a bar in PYVCP.
One way out, would probably be to use the inverse pins of our IO board ;to use a separate MUX8 and channel this to the visualisation bar.

Any help greatly appreciated.

kind regards,

Ghislain

Hello Andypugh,

thank you very much for the help. This works indeed.
For the sake of being complete: In the attachments you will find a copy of the XML file, an extract of the custom_postgui.hal file
As proof you can watch the workings on:https://youtu.be/95cdyaj2RBM

thanks once again,

Ghislain

Hello Andypugh,

I am aware of the experimental reverse-run initiative, although I must admit, we never tried it but it is definitely on our to do list.

tell a long story short:
I am the founder of the workshop at:http://www.vdwalle.com.
In 2012 we retrofitted a huge VMC; look at www.vdwalle.com/Norte
In 2015 we built a new plasma cutter; look at www.vdwalle.com/Cutty Spark
Both machines run on Ubuntu/LinuxCNC to our satifaction.
In this process all desktops became Linux.

We have an Agiecut 300 (a big iron of 1993) and a Charmilles Robofil190 (a small machine of 2003).
Since that time all brass wire machine builders have changed their process to under water cutting.
We particularly do not like this, because this type of machine becomes "unreachable for the operator" and look like a huge aquarium.
Sodick just made us an offer of 320.000 Euro for an aquarium of 3000 liter water to replace our Agie.

On the other hand there is a lot going on on the WEDM market in the far east.
They use molybdenum wire--rewinding their wire.
The nice thing of it; cost are going down and the speed is hugely going up; they claim speed x5 or even x10.
They are not that precise though; while a brass classic WEDM claims a precision in the 0.001 of a mm, Moly EDM is in the range of 0.01mm.
So I just want to see it for myself; the visum is applied for; once we have it, we fly over there and look for it ourselves.

On the other hand when I put an oscilloscope on our WEDM machines, talk to other people in the WEDM praxis, I hear a lot of fables and quasi truths.

Please have a look at:https://www.youtube.com/watch?v=kJNpy_HmPvw
When this Polish guy succeeds to run and "rape" this moly wire on LinuxCNC just using motion.enable, even running at 20mm/min
pulling and not breaking or shorting the wire, succes cannot be that far of.

So we started a small project building a generator and a gap detector to master the process in a trial setup, before building or retrofitting a machine. I am sure the reverse-run initiative will be integrated at its due time.

kind regards,

Ghislain

Hello Andypugh,

I have a question, motion.adaptive-feed related.
On our AGIE wire EDM we have beside the generator indicators and progressions bars an indication in mm/min how fast the wire progressed in the workpiece, updated every 5 seconds.
Just looking at the workpiece being eroded, the operator never can judge about the real progress the wire makes.
Visually a progress of 0.15mm/min is nothing different as a progress of 1.15mm/min.
Using this indicator, we adapt pulse voltage, pulse width, frequency and duty factor in the generator.
I have added a picture of the screen of the Agie in the attachment.

Talking LinuxCNC, how can this be built? Should I ask the programmers for it, or is it possible to use HAL and PYVCP or GladeVCP?

kind regards,